Selfish Mining

In ordinary proof-of-work operation, miners publish a discovered block so others can build on it, and rewards depend on which blocks remain in the accepted chain. A valid block can lose that reward opportunity if a competing branch wins.

Withholding changes what other miners can see: a group can work on a private branch while outside miners continue building on the visible chain, and selective publication can then cause competing work to become stale rather than part of the accepted history. The important distinction is strategic release.

An ordinary network delay or accidental competing block does not itself prove selfish mining, because the strategy uses information about private and public branch progress to improve the withholding group's expected reward outcome. Ittay Eyal and Emin Gun Sirer's paper, Majority is not Enough, analysed this incentive problem and showed conditions under which a colluding minority could earn more than its proportional resource share, challenging the idea that every profitable deviation requires majority control.

The analysis depends on its protocol and network assumptions, and computational share and the ability to attract other miners to a published branch affect the outcome. One model's threshold is not universal across networks.

Revenue advantage is also different from simply finding more blocks, since a group's block discoveries, accepted blocks and rewards can diverge when competing work is discarded, so evaluating the strategy requires the accepted-chain reward result rather than a count of all attempted work. Pool membership does not automatically imply selfish mining.

Pools normally combine resources to reduce reward variability and allocate proceeds among participants, so withholding is separate behaviour requiring evidence. In the paper's model, miners may have reason to join a group earning a disproportionate return, which is a modelled incentive and possible network consequence, not proof that every existing large pool has followed the strategy.

Selfish mining and a fifty-one-percent attack have different immediate descriptions: majority attacks are commonly discussed through transaction-history replacement or double spending, while selfish mining focuses on block publication and reward incentives and can be analysed below majority control. The existence of the strategy does not mean miners can forge another person's signature or spend arbitrary wallet balances, because protocol validity and cryptographic ownership checks still matter.

Consensus incentives do not confer unlimited asset control. Detection is difficult because public observers do not directly see every private branch while it is withheld.

Stale-block patterns, unusual publication timing and network evidence may support investigation, but none should be treated alone as conclusive proof, and one unusual event does not prove misconduct. For a non-finance manager assessing a mining venture or blockchain risk, ask what protocol assumptions support the claim, which reward evidence was measured, and whether the issue is a theoretical vulnerability, observed behaviour or the effect of a proposed defence, remembering that a research result is neither a guaranteed profitable trading plan nor a verdict on every network.